I want to use an additive hazards regression model (e.g. Aalen's model) instead of Cox regression, since my data shows non-proportional hazards.
Can anyone tell if there is a simple way to do this in SAS?
-Thanks.
Hello @SusanneWB ,
You can visit :
https://www.lexjansen.com/
(Lex Jansen is a former colleague of mine, consultant from SAS in the pharma industry)
(LexJansen site contains papers from proceedings from SAS conferences worldwide)
Then enter as search terms :
and hit enter.
You will get some useful papers from SAS conferences from the past.
See also the SAS documentation, ... for example these links :
Cheers,
Koen
Thanks a lot for the information. I will look at it.
Can I confirm then that Aalen's model and other cox alternatives have not yet been added to any SAS procedure, so I will need a macro?
Hello,
I cannot confirm that.
There are many alternatives to the COX PH model in the form of a procedure.
Or in the form of a SAS Studio task (behind these tasks, there are procedures of course).
See here :
SAS/STAT Software -- Focus area : Survival Analysis
https://support.sas.com/rnd/app/stat/procedures/SurvivalAnalysis.html
(remark the SAS/STAT QUANTLIFE Procedure is missing in the list)
SAS Studio 5.2 Task Reference
Nonparametric Survival Analysis
Cheers,
Koen
Thanks again.
I'm very interested in Aalen's additive hazard regression model. Can you also tell if this specific method is available in a procedure?
Hello,
The pure and full
<< Aalen's additive hazard regression model >>
is NOT available in a SAS procedure.
You will need a macro (there are some on internet).
Personally I would program it with SAS / IML (interactive matrix language).
Do you know the SAS matrix language? Very powerful and easy-to-use.
There's also a matrix language board under the Analytics header in these SAS communities if you need help using PROC IML (or the IML action set in PROC CAS).
Thanks,
Koen
Thanks a lot for the answer.
Then I will search for a macro or try with the SAS matrix language (which I have to get familiar with).
I think SAS should add alternative models for proportional hazards. We could open a ballot suggesting these other models:
- Aalen, O.O. 1989. A linear regression model for the analysis of life times. Statist. Med. 8: 907-925.
- Lin, D. Y. and Ying, Z. (1994). Semiparametric analysis of the additive risk model. Biometrika 81, 61–
71.
- J. Buckley, I. James, Linear regression with censored data, Biometrika 66 (1979) 429-436.
Maybe @StatDave knew it .
SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.